You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@harmony.apache.org by Mark Hindess <ma...@googlemail.com> on 2006/08/03 16:08:09 UTC
Re: [continuum] BUILD FAILURE: Classlib/win.ia32 Build/Test
On 3 August 2006 at 21:36, "Andrew Zhang" <zh...@gmail.com> wrote:
>
> On 8/3/06, Geir Magnusson Jr <ge...@pobox.com> wrote:
> >
> > Paulex Yang wrote:
> > > Andrew Zhang wrote:
> > >> The failure is caused by Harmony-1040.
> > >> The updated hostEquals(URL u1, URL u2) introduces a new bug:
> > >> It returns true when u1.host = null and u2.host = "" while it should
> > >> return
> > >> false.
> > >
> > >> Sorry for the mistake caused by my patch. I'll update a new patch to
> > >> solve this problem.
Ideally we should also add a regression test to luni to catch this
problem. The fact that the problem wasn't caught means the current
luni don't have sufficient coverage so adding a test to luni would
help improve this.
I think we should always try to do this when a change in a module
passes that modules tests but fails another modules tests (as long
as it is reasonably practical to do so).
Regards,
Mark.
> > > Great, thank you to find this, Andrew. And don't need to say sorry, it's
> > > me who takes responsibility on this and then owes the beer:-(.
> >
> > Out of curiosity, did either of you run the test suite on windows before
> > offering the patch or committing it? This isn't about "blame" or such,
> > but about learning about how to make a better process....
>
>
> I did run luni and nio module test on both windows and linux, but security
> module fails...
>
> It's a lesson for me. I'll pay more attention next time.
>
>
>
> > There's no harm done - it was found fast, and fixed fast, so kudos to
> > you guys for working together on that.
> >
> > geir
---------------------------------------------------------------------
Terms of use : http://incubator.apache.org/harmony/mailing.html
To unsubscribe, e-mail: harmony-dev-unsubscribe@incubator.apache.org
For additional commands, e-mail: harmony-dev-help@incubator.apache.org
Re: [continuum] BUILD FAILURE: Classlib/win.ia32 Build/Test
Posted by Geir Magnusson Jr <ge...@pobox.com>.
Mark Hindess wrote:
> On 3 August 2006 at 21:36, "Andrew Zhang" <zh...@gmail.com> wrote:
>> On 8/3/06, Geir Magnusson Jr <ge...@pobox.com> wrote:
>>> Paulex Yang wrote:
>>>> Andrew Zhang wrote:
>>>>> The failure is caused by Harmony-1040.
>>>>> The updated hostEquals(URL u1, URL u2) introduces a new bug:
>>>>> It returns true when u1.host = null and u2.host = "" while it should
>>>>> return
>>>>> false.
>>>>> Sorry for the mistake caused by my patch. I'll update a new patch to
>>>>> solve this problem.
>
> Ideally we should also add a regression test to luni to catch this
> problem. The fact that the problem wasn't caught means the current
> luni don't have sufficient coverage so adding a test to luni would
> help improve this.
Whatever happened to "module + 1 dependency level" for testing?
>
> I think we should always try to do this when a change in a module
> passes that modules tests but fails another modules tests (as long
> as it is reasonably practical to do so).
>
> Regards,
> Mark.
>
>>>> Great, thank you to find this, Andrew. And don't need to say sorry, it's
>>>> me who takes responsibility on this and then owes the beer:-(.
>>> Out of curiosity, did either of you run the test suite on windows before
>>> offering the patch or committing it? This isn't about "blame" or such,
>>> but about learning about how to make a better process....
>>
>> I did run luni and nio module test on both windows and linux, but security
>> module fails...
>>
>> It's a lesson for me. I'll pay more attention next time.
>>
>>
>>
>>> There's no harm done - it was found fast, and fixed fast, so kudos to
>>> you guys for working together on that.
>>>
>>> geir
>
>
> ---------------------------------------------------------------------
> Terms of use : http://incubator.apache.org/harmony/mailing.html
> To unsubscribe, e-mail: harmony-dev-unsubscribe@incubator.apache.org
> For additional commands, e-mail: harmony-dev-help@incubator.apache.org
>
>
>
---------------------------------------------------------------------
Terms of use : http://incubator.apache.org/harmony/mailing.html
To unsubscribe, e-mail: harmony-dev-unsubscribe@incubator.apache.org
For additional commands, e-mail: harmony-dev-help@incubator.apache.org
Re: [continuum] BUILD FAILURE: Classlib/win.ia32 Build/Test
Posted by Stepan Mishura <st...@gmail.com>.
On 8/3/06, Paulex Yang wrote:
>
> Mark Hindess wrote:
> > On 3 August 2006 at 21:36, "Andrew Zhang" <zh...@gmail.com>
> wrote:
> >
> >> On 8/3/06, Geir Magnusson Jr <ge...@pobox.com> wrote:
> >>
> >>> Paulex Yang wrote:
> >>>
> >>>> Andrew Zhang wrote:
> >>>>
> >>>>> The failure is caused by Harmony-1040.
> >>>>> The updated hostEquals(URL u1, URL u2) introduces a new bug:
> >>>>> It returns true when u1.host = null and u2.host = "" while it should
> >>>>> return
> >>>>> false.
> >>>>>
> >>>>> Sorry for the mistake caused by my patch. I'll update a new patch to
> >>>>> solve this problem.
> >>>>>
> >
> > Ideally we should also add a regression test to luni to catch this
> > problem. The fact that the problem wasn't caught means the current
> > luni don't have sufficient coverage so adding a test to luni would
> > help improve this.
> >
> > I think we should always try to do this when a change in a module
> > passes that modules tests but fails another modules tests (as long
> > as it is reasonably practical to do so).
> >
> Agree, I expect Andrew will include this in his new patch:).
>
> The problem more serious is how we prevent these things before commit.
IMHO if a patch fixes 'linu' module it would be better to run all tests.
Thanks,
Stepan.
> Regards,
> > Mark.
> >
> >
> >>>> Great, thank you to find this, Andrew. And don't need to say sorry,
> it's
> >>>> me who takes responsibility on this and then owes the beer:-(.
> >>>>
> >>> Out of curiosity, did either of you run the test suite on windows
> before
> >>> offering the patch or committing it? This isn't about "blame" or
> such,
> >>> but about learning about how to make a better process....
> >>>
> >> I did run luni and nio module test on both windows and linux, but
> security
> >> module fails...
> >>
> >> It's a lesson for me. I'll pay more attention next time.
> >>
> >>
> >>
> >>
> >>> There's no harm done - it was found fast, and fixed fast, so kudos to
> >>> you guys for working together on that.
> >>>
> >>> geir
> >>>
> >
>
>
--
Thanks,
Stepan Mishura
Intel Middleware Products Division
------------------------------------------------------
Terms of use : http://incubator.apache.org/harmony/mailing.html
To unsubscribe, e-mail: harmony-dev-unsubscribe@incubator.apache.org
For additional commands, e-mail: harmony-dev-help@incubator.apache.org
Re: [continuum] BUILD FAILURE: Classlib/win.ia32 Build/Test
Posted by Paulex Yang <pa...@gmail.com>.
Mark Hindess wrote:
> On 3 August 2006 at 21:36, "Andrew Zhang" <zh...@gmail.com> wrote:
>
>> On 8/3/06, Geir Magnusson Jr <ge...@pobox.com> wrote:
>>
>>> Paulex Yang wrote:
>>>
>>>> Andrew Zhang wrote:
>>>>
>>>>> The failure is caused by Harmony-1040.
>>>>> The updated hostEquals(URL u1, URL u2) introduces a new bug:
>>>>> It returns true when u1.host = null and u2.host = "" while it should
>>>>> return
>>>>> false.
>>>>>
>>>>> Sorry for the mistake caused by my patch. I'll update a new patch to
>>>>> solve this problem.
>>>>>
>
> Ideally we should also add a regression test to luni to catch this
> problem. The fact that the problem wasn't caught means the current
> luni don't have sufficient coverage so adding a test to luni would
> help improve this.
>
> I think we should always try to do this when a change in a module
> passes that modules tests but fails another modules tests (as long
> as it is reasonably practical to do so).
>
Agree, I expect Andrew will include this in his new patch:).
The problem more serious is how we prevent these things before commit.
> Regards,
> Mark.
>
>
>>>> Great, thank you to find this, Andrew. And don't need to say sorry, it's
>>>> me who takes responsibility on this and then owes the beer:-(.
>>>>
>>> Out of curiosity, did either of you run the test suite on windows before
>>> offering the patch or committing it? This isn't about "blame" or such,
>>> but about learning about how to make a better process....
>>>
>> I did run luni and nio module test on both windows and linux, but security
>> module fails...
>>
>> It's a lesson for me. I'll pay more attention next time.
>>
>>
>>
>>
>>> There's no harm done - it was found fast, and fixed fast, so kudos to
>>> you guys for working together on that.
>>>
>>> geir
>>>
>
>
> ---------------------------------------------------------------------
> Terms of use : http://incubator.apache.org/harmony/mailing.html
> To unsubscribe, e-mail: harmony-dev-unsubscribe@incubator.apache.org
> For additional commands, e-mail: harmony-dev-help@incubator.apache.org
>
>
>
--
Paulex Yang
China Software Development Lab
IBM
---------------------------------------------------------------------
Terms of use : http://incubator.apache.org/harmony/mailing.html
To unsubscribe, e-mail: harmony-dev-unsubscribe@incubator.apache.org
For additional commands, e-mail: harmony-dev-help@incubator.apache.org